Keyoku-ai

Keyoku-ai / keyoku

Public

Persistent AI memory for OpenClaw — auto-recall, auto-capture, heartbeat, knowledge graph

www.npmjs.compackage@keyokuopenclaw agents ai ai-memory knowledge-graph memory
12
4
100% credibility
Found Mar 16, 2026 at 12 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

Keyoku is a plugin that adds persistent, intelligent memory to OpenClaw AI assistants, enabling automatic recall, fact extraction, knowledge graphs, and proactive heartbeat signals.

How It Works

1
🔍 Hear about super memory for your AI helper

You're chatting with your AI assistant and wish it remembered your preferences and past talks without repeating yourself.

2
🚀 Add memory with one easy command

Run a simple setup command that finds your AI helper and turns on powerful remembering abilities automatically.

3
⚙️ It sets everything up for you

The magic connects your helper to a smart memory system that learns from every conversation and keeps facts organized.

4
🔄 Restart your AI assistant

Close and reopen your helper, and now it knows you deeply from the start.

5
🧠 Chat and see instant smarts

Talk naturally, and your assistant recalls details, preferences, and plans without you reminding it.

6
⏰ Get helpful nudges automatically

Your helper checks in proactively about deadlines, changes, or unfinished ideas at the right times.

🎉 Enjoy a truly personal AI partner

Over time, your assistant acts like a thoughtful friend who remembers everything and helps without being asked.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 12 to 12 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is keyoku?

Keyoku brings persistent memory to OpenClaw AI agents via a TypeScript plugin, storing facts in local SQLite with HNSW vectors for fast semantic search. It auto-captures conversation insights, recalls relevant memories before responses, and runs a heartbeat scanner for deadlines, conflicts, and trends using a knowledge graph. Developers get github persistent storage that mimics human recall, with tools for scheduling and stats.

Why is it gaining traction?

It upgrades OpenClaw's flat MEMORY.md files to smart persistent memory llm: three-tier dedup, Ebbinghaus decay, conflict resolution, and proactive signals like sentiment shifts or stalled goals. The heartbeat delivers structured briefs at observe/suggest/act levels, far beyond static checks. One-command npx init with auto-migration hooks users instantly, plus a standalone client for custom agents.

Who should use this?

OpenClaw builders creating task-tracking assistants that remember preferences and nudge on deadlines. Solo devs using chat for project management, or teams in Slack/Discord needing auto-capture of decisions and persistent memory programming across sessions. Pairs well with persistent memory intel for local AI workflows.

Verdict

Promising for OpenClaw users wanting persistent memory pmem without clouds (12 stars, 1.0% credibility signals early stage), backed by thorough docs and CLI like "openclaw memory stats." Try the init script; it'll mature fast if adoption grows.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.